[Android docs] Add building instructions and code snippet

docs/source/using-executorch-android.md

@@ -68,17 +68,85 @@ dependencies {
 
 Now you can compile your app with the ExecuTorch Android library.
 
-### Building from Source
+## Building from Source
 
-TODO Instructions on re-creating and customizing the Android AAR.
+`build/build_android_library.sh` is a helper script to build the Java library (into .jar), native library (into .so), and the packaged AAR file. It can also build
+demo apps to showcase the AAR is integrated into a user app correctly.
+
+You need Android [SDK](https://developer.android.com/studio) and [NDK](https://developer.android.com/ndk/downloads) to use it.
+
+Current NDK version used in ExecuTorch CI: r27b.
+
+You need to set `ANDROID_NDK` to the correct NDK root (containing NOTICE file).
+
+```
+export ANDROID_NDK=/path/to/ndk
+sh build/build_android_library.sh
+```
+
+### Optional environment variables
+
+Optionally, set these environment variables before running `build_android_library.sh`.
+
+#### ANDROID_ABIS
+Set environment variable `ANDROID_ABIS` to either `arm64-v8a` or `x86_64` if you only need to build the native library for one ABI only.
+```
+export ANDROID_ABIS=arm64-v8a
+# or
+# export ANDROID_ABIS=x86_64
+sh build/build_android_library.sh
+```
+
+#### EXECUTORCH_CMAKE_BUILD_TYPE
+Set environment variable `EXECUTORCH_CMAKE_BUILD_TYPE` to `Release` or `Debug` based on your needs.
+
+#### Using MediaTek backend
+
+To use [MediaTek backend](https://pytorch.org/executorch/main/backends-mediatek.html),
+after installing and setting up the SDK, set `NEURON_BUFFER_ALLOCATOR_LIB` and `NEURON_USDK_ADAPTER_LIB` to the corresponding path.
+
+#### Using Qualcomm AI Engine Backend
+
+To use [Qualcomm AI Engine Backend](https://pytorch.org/executorch/main/backends-qualcomm.html#qualcomm-ai-engine-backend),
+after installing and setting up the SDK, set `QNN_SDK_ROOT` to the corresponding path
 
 ## Android Backends
 
 TODO Describe commonly used backends, including XNN, Vulkan, and NPUs.
 
 ## Runtime Integration
 
-TODO Code sample in Java
+Here is an example code sample in Java that demonstrates how to integrate ExecuTorch into an Android app:
+
+```java
+import org.pytorch.executorch.EValue;
+import org.pytorch.executorch.Module;
+import org.pytorch.executorch.Tensor;
+
+public class MainActivity extends Activity {
+    private Module module;
+
+    @Override
+    protected void onCreate(Bundle savedInstanceState) {
+        super.onCreate(savedInstanceState);
+        // Load the ExecuTorch module
+        module = Module.load("/path/to/module.pte");
+    }
+    public void runInference(View view) {
+        // Prepare input data
+        Tensor input = Tensor.fromBlob(getInputData());
+        // Run inference
+        Tensor output = module.forward(EValue.from(input))[0].toTensor();
+        // Process output data
+        processOutput(output);
+    }
+}
+```
+This example loads an ExecuTorch module, prepares input data, runs inference, and processes the output data.
+
+Please use [ExecuTorchDemo](https://github.com/pytorch/executorch/tree/main/examples/demo-apps/android/ExecuTorchDemo)
+and [LlamaDemo](https://github.com/pytorch/executorch/tree/main/examples/demo-apps/android/LlamaDemo) for the code examples
+using ExecuTorch AAR package.
 
 ## Next Steps